Selection of measures suitable for evaluating change in fibromyalgia clinical trials.

Objective: To use growth curve model methods to investigate which of 46 variables associated with fibromyalgia (FM) clinical trials are potentially useful in measuring change.

Methods: For each of the 46 variables the reliability of change and corresponding standard error were estimated. Data were from a randomized clinical trial designed to compare the effectiveness of biofeedback/relaxation, exercise, and a combined program for the treatment of FM. There was also a control group.

Results: The reliabilities of change for the outcome variables Myalgic Score, Tenderpoint Count, and Tenderpoint Index, as well as the variable, number of words chosen from the McGill Pain Questionnaire were quite acceptable for both the 18 and 58 week time periods. There were 9 other variables that had reliabilities of change in the acceptable range.

Conclusion: Thirteen of 46 variables considered had desirable reliabilities of change by the methods used. The 3 variables often used as measures of disease activity as well as one pain variable were among the 13. Data from this study produced no acceptable reliabilities of change for the 33 other variables. For these variables, further evaluation is required.